The flexibility and instantaneity of existing water environment warning system seems difficult to satisfy the increasingly severe water environment security situation and water pollution events. A dynamic early warning system structure based on pollutant dispersion simulation has been put forward, with some research on the construction of pollutant dispersion model, dynamic trigger optimization model , dynamic display of simulation results and other key techniques in detail. Take the Qiantang River in Hangzhou section for example, this paper builds the water hydrodynamic model and the convection diffusion model, simulates the dispersion of assumed pollutants and then displays the fuctions of providing dynamic early warning services.The application shows that this system can provide accurate and convenient warning of the influence range and degree of water pollutants. It also provides decision support for relevant government departments, water companies and other enterprises to take emergency measures timely and effectively.

In order to meet the requirement to quickly solve the water quality equation for an unexpected water pollution incident, this dissertation, based on the paralleled algorithm under Java Parallel Processing Framework (JPPF), does research about the process to dynamically decomposes the iterative process into calculation tasks and distribute the tasks by the API of JPPF to the paralleled nodes for calculation. And the simulation result of one-dimension water quality equation shows that parallel computing method could reduce the time complexity from to , not only resulting in a significant improvement in calculation speed, but in a higher reliability and stability.